0:006:12

Recursos

Resumen

Take a proof that torus packets in SL_n(Z)\SL_n(R) equidistribute when n is a prime at least five. For each step, say which way the limit could fail that step rules out, and what the step hands to the next one.

The claim

Each ideal class, embedded as a lattice with covolume one, gives a compact diagonal orbit together with its sign translates. The unit theorem is what closes the loop. Weight the orbits of a packet by volume and you get a measure μ. As D(M) → ∞, μ converges weakly to the Haar probability m_n and no mass escapes [1:00]. The claim holds separately for each local homothety type. It says nothing about single orbits, rates or composite degrees [1:24].

The chain from counting to rigidity [1:38]

  1. Unfold [1:54]: the packet average of a lattice-vector count becomes (1/(√D κ_K)) Σ P(𝔞) V_f(N𝔞/Q). Here P(𝔞) is the local containment weight and Q = q√D.
  2. Bound [2:50]: the short-interval count is ≪ κ_K y / q. κ_K cancels, so upper bounds are enough and no subconvexity is needed. Shiu's theorem and Stark's zero-free region give the bound. The exceptional zero is ruled out because it would force a quadratic subfield, and odd prime degree has none.
  3. Two uses [3:50]: a cube at the origin gives tightness, so no escape of mass. A box around a vector with nonzero coordinates gives μ(xB(r)) ≤ C rⁿ.
  4. Entropy [4:21]: about r^{-(n-1)} balls cover a two-sided tube, and each carries at most rⁿ. One factor r = e^{-λt} is left, so entropy is ≥ λ/3. This holds for every ξ ≤ cμ, which rules out zero-entropy components.
  5. Classify [5:22]: by Einsiedler–Katok–Lindenstrauss, positive-entropy ergodic components are Haar, so μ = m_n.

Reading rule: for each estimate, ask which failure it removes: mass leaking into the cusp, mass piling onto small balls, or zero-entropy pieces.

Equidistribution of prime-degree torus packets

OpenAI Math Preprints

Clase 34 / 100

Ver clases